I'm currently finishing a Master of IT.
Featured Projects
Computer Vision Sports Analytics (badminton stroke classifier) - Extends the state-of-the-art BST CV + time-series architecture, recording new top benchmarks on the published ShuttleSet evaluation set.
- Part of a group-project repo (currently WIP). The src/bst_refactor/ subtree is my contribution, covering model build and data pipeline.
- Currently in build: a new model layer to overcome the feature signal ceiling, adding RGB input from a fine-tuned X3D-S backbone.
Object Detection (Australian Ferns) - Comparative study of two object detection models on a hand-annotated dataset.
- Dataset scraped from iNaturalist and annotated in CVAT; splits stratified by per-class annotation density. The pipeline is reproducible from the source CSVs.
- Compares YOLOv8m (CNN-based detector) with RetinaNet on a MiT-b0 backbone (hybrid CNN+transformer).
- Explainable-AI tools (GradCAM, EigenCAM) are used to understand classification confusion, introduce hard negative images, validate improvements in feature mapping.
- Each architecture was refined across a series of hyperparameter grids.
LSTM Market Classifier - Time-series inference model and data pipeline.
- Analytical write-up, Python scraper modules and detailed notebook.
- Most of the implementation code is private, but shoot me a message if you're working on something similar and want to talk about it.
Other projects include a Java Chess GUI, Java Ordering System and Relational Database with Custom Views.